Water quality monitoring unmanned ship system and method based on multi-sensor fusion

Abstract

The invention relates to the technical field of unmanned ships, and particularly discloses a water quality monitoring unmanned ship system and method based on multi-sensor fusion, and the system is characterized in that an initial monitoring module obtains the azimuth information of all environmental obstacles and various types of water quality monitoring data of an unmanned ship based on various sensors; the map stroking module is used for positioning the azimuth information of all environmental obstacles obtained in a preset period according to a time sequence and stroking in an electronic map to obtain a current water area boundary and a dynamic track of a dynamic obstacle; the path optimization module is used for generating an optimized sampling path in real time based on the local mutation range of all types of water quality monitoring data acquired in a preset period, the current water area boundary and the dynamic track of the dynamic obstacle; the monitoring and returning module controls the unmanned ship to sail on the water area according to the optimized sampling path and returns various types of water quality monitoring data collected by various sensors in real time; the efficient planning of the sampling path of the water quality monitoring unmanned ship and the efficient monitoring of the water quality of the water area are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned vessels, and particularly to a water quality monitoring unmanned vessel system and method based on multi-sensor fusion. Background Art

[0002] In the fields of water resource protection and water environment management, water quality monitoring plays a crucial role. With the rapid development of industrialization and urbanization, water resources are facing increasingly complex pollution problems. Accurately and comprehensively monitoring the water quality status is essential for protecting water resources, maintaining ecological balance, and ensuring human health. Traditional water quality monitoring methods mostly rely on manual sampling and fixed monitoring stations, which have disadvantages such as limited monitoring range, low spatio-temporal resolution, and inability to obtain data in real time. Water quality monitoring unmanned vessel systems and methods have emerged, bringing new breakthroughs to water quality monitoring. Such systems use unmanned vessels as mobile monitoring platforms, which can not only improve the efficiency and accuracy of water quality monitoring, but also reduce labor costs and adapt to complex and changeable water environments. With the continuous progress of technology, water quality monitoring unmanned vessel systems have broad application prospects in the fields of water resource protection, water environment monitoring, etc., and are expected to become an important means of future water quality monitoring.

[0003] However, the existing water quality monitoring technologies have many deficiencies. During the monitoring process, it is difficult to accurately obtain the dynamic trajectories of the current water area boundary and dynamic obstacles, which is not conducive to the overall grasp of the water environment. Due to the lack of comprehensive analysis of water quality monitoring data and environmental information, the sampling path lacks scientificity and flexibility, affecting the timely acquisition and analysis of data, and thus unable to provide timely and accurate decision-making basis for water resource management.

[0004] Therefore, the present invention proposes a water quality monitoring unmanned vessel system and method based on multi-sensor fusion. Summary of the Invention

[0005] The present invention provides a water quality monitoring unmanned vessel system and method based on multi-sensor fusion. The system is equipped with a variety of sensors and can autonomously navigate in the water area and collect various types of water quality monitoring data in real time. Through multi-sensor fusion technology, the data of different types of sensors can be comprehensively analyzed to improve the accuracy and reliability of monitoring. At the same time, the unmanned vessel can flexibly adjust the sampling path according to the real-time obtained environmental information to achieve comprehensive and dynamic monitoring of the water area.

[0006] The present invention provides a water quality monitoring unmanned vessel system based on multi-sensor fusion, including: An initial monitoring module, configured to obtain the azimuth information of all environmental obstacles and various types of water quality monitoring data in real time based on a variety of sensors when the unmanned vessel sails on the water area according to the initial sampling path; A map outlining module, which is used to perform chronological positioning on the orientation information of all environmental obstacles obtained within a preset period and outline them on an electronic map to obtain the current water area boundary and the dynamic trajectories of dynamic obstacles; A path optimization module, which is used to generate an optimized sampling path in real time based on the local mutation range of all types of water quality monitoring data obtained within a preset period, the current water area boundary, and the dynamic trajectories of dynamic obstacles; A monitoring and feedback module, which is used to control the unmanned ship to sail on the water area according to the optimized sampling path and feedback the multi-type water quality monitoring data collected in real time by a variety of sensors.

[0007] Preferably, the initial monitoring module includes: An obstacle detection sub-module, which is used to obtain the distances of all environmental obstacles when the unmanned ship sails on the water area according to the initial sampling path based on lidar in real time, and obtain the self-positioning information of the unmanned ship when it sails on the water area according to the initial sampling path based on the GPS module; A water quality monitoring sub-module, which is used to obtain multi-type water quality monitoring data when the unmanned ship sails on the water area according to the initial sampling path based on multi-type water quality sensors.

[0008] Preferably, it further includes: An underwater camera and a network camera are used to obtain and feedback an underwater monitoring video and an above-water monitoring video respectively.

[0009] Preferably, the map outlining module includes: An obstacle boundary point marking sub-module, which is used to mark on the electronic map based on the direction and distance in the orientation information of all environmental obstacles obtained at each moment within a preset period and the self-positioning information of the unmanned ship to obtain the boundary marking points at each moment within the preset period; A determined boundary calibration sub-module, which is used to count the marking frequencies of each boundary marking point at all moments within a preset period, screen out all boundary marking points with a marking frequency greater than a preset marking frequency threshold among all boundary marking points within the preset period as determined boundary points, determine the determined boundary based on all determined boundary points, and regard all the remaining boundary marking points except the determined boundary points among all boundary marking points at all moments as the first undetermined points; A dynamic trajectory calculation sub-module, which is used to determine the dynamic trajectories of all dynamic obstacles based on all the first undetermined points and the corresponding calibration times; A water area boundary outlining sub-module, which is used to determine the current water area boundary based on the dynamic trajectories of all dynamic obstacles and the determined boundary.

[0010] Preferably, the dynamic trajectory calculation sub-module includes: a hypothetical path calibration unit, configured to aggregate all first undetermined points at each moment within a preset period to obtain a first undetermined point set at each moment, select any first undetermined point from the first undetermined point set at each moment within the preset period and aggregate the points to obtain all first hypothetical target drift paths, and use, among all the first hypothetical target drift paths, the hypothetical target drift path that intersects the determined boundary at most once as the second hypothetical target drift path; a speed stability evaluation unit, configured to calculate a drift speed stability of each second assumed target drift path; A relative position analysis unit, configured to determine a boundary distance factor, a crossing number penalty factor, and a direction consistency factor of each second hypothetical target drift path; an occurrence probability evaluation unit, configured to evaluate the occurrence probability of each hypothetical target drift path based on a boundary distance factor, a crossing number penalty factor, and a direction consistency factor of each second hypothetical target drift path; The rationality determination unit is used to calculate the rationality of each assumed target drift probability based on the drift speed stability and occurrence probability of each second assumed target drift path, and regard all second assumed target drift paths with a rationality greater than a preset rationality threshold as the dynamic trajectories of all dynamic obstacles.

[0011] Preferably, the relative position analysis unit includes: a boundary distance factor determination subunit, configured to fit a path line of each second hypothetical target drift path, and determine a boundary distance factor corresponding to the second hypothetical target drift path based on the distances between all position points and corresponding distance statistical points in each path line; a crossing number penalty factor determination subunit, configured to define a calculation formula for the crossing number penalty factor based on an exponential function, and determine the crossing number penalty factor for each second hypothetical target drift path based on the number of intersections between each second hypothetical target drift path and the predetermined boundary and the calculation formula for the crossing number penalty factor; The direction consistency factor determination subunit is used to take the similarity between the path tangent angle vector and the boundary tangent vector of each second hypothetical target drift path as the direction consistency factor.

[0012] Preferably, the water area boundary stroke submodule includes: An undetermined point definition unit is used to treat all boundary marking points at all times, except for all boundary marking points covered by the dynamic trajectories of the determined boundary and all dynamic obstacles, as second undetermined points; A water area boundary stroking unit is used to perform clustering analysis on the second undetermined points with the distance between all the second undetermined points as the clustering condition, determine multiple clustering position points, and connect the determined boundary with all the clustering position points on the premise of avoiding intersection with the dynamic trajectories of all dynamic obstacles to obtain the current water area boundary.

[0013] Preferably, the process of the path optimization module determining the local mutation range of all types of water quality monitoring data obtained within a preset period includes: Based on the spatial correlation constraint effect between the current local mutation boundary and the current water area boundary of all types of water quality monitoring data obtained within a preset period, the synergy effect between all types of water quality monitoring data, the trajectory correlation effect of dynamic obstacles, and a preset pollution source mutation mode library, analyze the mutation cause of the current local mutation boundary of each type of water quality monitoring data; Based on the mutation cause of the current local mutation boundary of each type of water quality monitoring data and the current water area boundary, perform continuous extrapolation on the current local mutation boundary of each type of water quality monitoring data to obtain the local mutation range of each type of water quality monitoring data.

[0014] Preferably, the path optimization module includes: An objective function generation sub-module is used to generate an objective function based on the pollution coverage, voyage cost, and obstacle avoidance safety distance; A sampling path optimization sub-module is used to generate an optimized sampling path in real time with the goal of minimizing the output value of the objective function.

[0015] The present invention provides a water quality monitoring method based on multi-sensor fusion, including: Based on multiple sensors, obtain the azimuth information of all environmental obstacles and multiple types of water quality monitoring data in real time when the unmanned ship sails on the water area according to the initial sampling path; Perform chronological positioning on the azimuth information of all environmental obstacles obtained within a preset period and stroke them on the electronic map to obtain the current water area boundary and the dynamic trajectories of dynamic obstacles; Based on the local mutation range of all types of water quality monitoring data obtained within a preset period, the current water area boundary, and the dynamic trajectories of dynamic obstacles, generate an optimized sampling path in real time; Control the unmanned ship to sail on the water area according to the optimized sampling path and transmit back the multiple types of water quality monitoring data collected in real time by using multiple sensors.

[0016] The beneficial effects of the present invention compared with the prior art are as follows: By using a variety of sensors to real-time obtain the azimuth information of environmental obstacles and various types of water quality monitoring data during the navigation of the unmanned ship, it provides basic data support for comprehensively understanding the water area conditions subsequently, ensuring the integrity and timeliness of the information. The azimuth information of environmental obstacles is located in chronological order and outlined on the electronic map, accurately obtaining the water area boundary and the dynamic trajectory of dynamic obstacles, visually presenting the changes in the water area environment, providing a visual basis for the navigation path planning, and effectively ensuring the navigation safety of the unmanned ship. According to the local mutation range of the water quality monitoring data, as well as the water area boundary and the trajectory of dynamic obstacles, an optimized sampling path is generated in real-time, taking into account both the key areas of water quality monitoring and navigation safety, making the sampling more targeted and scientific, improving the monitoring efficiency and data accuracy. Control the unmanned ship to navigate according to the optimized path and transmit back the real-time collected water quality monitoring data, ensuring that the data can be transmitted in time for analysis and research, realizing the dynamic and efficient monitoring of the water quality of the water area, and assisting in the formulation of water resource management and protection decisions.

[0017] Other features and advantages of the present invention will be described in the following specification, and, in part, will be obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the structures specifically pointed out in this application document.

[0018] The technical solutions of the present invention will be further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, and do not constitute a limitation to the present invention. In the drawings: Figure 1 is a schematic diagram of a water quality monitoring unmanned ship system based on multi-sensor fusion in an embodiment of the present invention; Figure 2 is a schematic diagram of an initial monitoring module in an embodiment of the present invention; Figure 3 is a schematic diagram of the initial calibration process of the water area boundary in an embodiment of the present invention; Figure 4 is a schematic diagram of the sampling path optimization process in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] The following describes the preferred embodiments of the present invention with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0021] As Figure 1 shown, the present invention provides an embodiment of a water quality monitoring unmanned ship system based on multi-sensor fusion, including: Initial monitoring module, which is used to obtain the azimuth information of all environmental obstacles and various types of water quality monitoring data in real time based on multiple sensors when the unmanned ship sails on the water along the initial sampling path; Map outlining module, which is used to locate the azimuth information of all environmental obstacles obtained within a preset period in chronological order and outline them on the electronic map to obtain the current water area boundary and the dynamic trajectories of dynamic obstacles; Path optimization module, which is used to generate an optimized sampling path in real time based on the local mutation range of all types of water quality monitoring data obtained within a preset period and the current water area boundary and the dynamic trajectories of dynamic obstacles; Monitoring and transmission module, which is used to control the unmanned ship to sail on the water along the optimized sampling path and transmit various types of water quality monitoring data collected in real time by multiple sensors to the cloud or the monitoring center.

[0022] In this embodiment, the initial sampling path is a pre-set sailing route when the unmanned ship starts the water quality monitoring task, providing a starting trajectory reference for subsequent monitoring. For example, when monitoring a regular square water area, the initial sampling path may be set to start from a corner of the water area and traverse the water area back and forth in a "zigzag" shape.

[0023] In this embodiment, the azimuth information of environmental obstacles refers to the information such as the direction and distance of surrounding obstacles relative to itself obtained in real time by the sensors during the sailing process of the unmanned ship, helping the unmanned ship perceive the surrounding environmental conditions. For example, the unmanned ship learns through lidar and GPS that there is a bridge 50 meters ahead and 30 degrees to the left.

[0024] In this embodiment, various types of water quality monitoring data are data reflecting different aspects of water quality characteristics collected in the water area by various water quality sensors. Data such as pH value, dissolved oxygen content, and heavy metal concentration obtained through different sensors together constitute various types of water quality monitoring data.

[0025] In this embodiment, the preset period is a fixed duration set artificially. During this time period, the unmanned ship continuously collects environmental and water quality-related data for subsequent centralized analysis and processing. For example, the preset period is set to 30 minutes, and within this half hour, the unmanned ship continuously collects various types of data.

[0026] In this embodiment, the current water area boundary is the water area range boundary determined on the electronic map through processing and analysis based on the azimuth information of environmental obstacles within a preset period, delimiting a safe area for the unmanned ship to sail. For example, when monitoring a river, the river boundary range determined in combination with the positions of fixed obstacles on the shore and in the water.

[0027] In this embodiment, the dynamic trajectory of a dynamic obstacle refers to the moving route of the obstacle in the water area obtained by calculation according to the change of the azimuth information of the dynamic obstacle within a preset period, so as to enable the unmanned ship to plan an avoidance path in advance. For example, the moving trajectory of other ships sailing in the water area, which is determined by their azimuth information at different times.

[0028] In this embodiment, the local mutation range of all types of water quality monitoring data is the range of areas where water quality mutations may occur, which is deduced by comprehensively considering the cooperation between water quality data, the spatial association with the water area boundary, the trajectory association of dynamic obstacles, and the pollution source mutation pattern library. For example, when multiple water quality indicators in a certain area are abnormal at the same time, the possible affected water area range is deduced by combining the surrounding environmental factors.

[0029] In this embodiment, the optimized sampling path is the navigation path of the unmanned ship generated in real time by the path optimization module based on the pollution coverage, voyage cost, obstacle avoidance safety distance, etc., with the goal of minimizing the function value, and in combination with the water quality data mutation range, water area boundary, and dynamic obstacle trajectory, so as to improve the monitoring efficiency and accuracy. For example, a tortuous but efficient navigation route that avoids obstacles and preferentially covers areas that may be polluted.

[0030] As Figure 2 shown, further, in order to more accurately obtain the distance of environmental obstacles, the self-positioning information, and various types of water quality monitoring data when the unmanned ship is sailing, an initial monitoring module is proposed, including: An obstacle detection sub-module, which is used to obtain the distances of all environmental obstacles when the unmanned ship is sailing on the water area along the initial sampling path based on lidar in real time, and obtain the self-positioning information of the unmanned ship when it is sailing on the water area along the initial sampling path based on the GPS module in real time; A water quality monitoring sub-module, which is used to obtain various types of water quality monitoring data when the unmanned ship is sailing on the water area along the initial sampling path based on various water quality sensors.

[0031] In this embodiment, the distance of all environmental obstacles is obtained in real time based on the lidar when the unmanned ship sails on the water along the initial sampling path, and the self-positioning information of the unmanned ship when it sails on the water along the initial sampling path is obtained in real time based on the GPS module. This means using the principle of the lidar emitting laser beams and receiving reflected light to measure in real time the distance between the unmanned ship and all surrounding environmental obstacles when sailing along the initial sampling path. At the same time, with the help of the GPS module, the position information of the unmanned ship when sailing along the initial sampling path in this water area is determined in real time. These two sets of information play a key role in enabling the unmanned ship to perceive the surrounding environment and plan a safe and effective navigation route. For example, when the unmanned ship moves forward along the initial sampling path on a river, the lidar can measure that there is a big rock 20 meters ahead and a floating object 15 meters at the right rear. At the same time, the GPS module real-time feedbacks that the unmanned ship is at a specific longitude and latitude position on the river, enabling the unmanned ship to clearly know its relative position to these obstacles so as to make reasonable navigation decisions.

[0032] Furthermore, in order to provide underwater and water surface visual monitoring data, it is proposed to add a bottom camera and a network camera to obtain and transmit back underwater and water surface monitoring videos, including: The bottom camera and the network camera are used to obtain and transmit back underwater monitoring videos and water surface monitoring videos respectively.

[0033] In this embodiment, the bottom camera and the network camera are used to obtain and transmit back underwater monitoring videos and water surface monitoring videos respectively. This means that in the unmanned ship system, the bottom camera is responsible for shooting the real-time picture of the water area under the unmanned ship to form an underwater monitoring video, and the network camera focuses on the scene above the water surface for shooting to generate a water surface monitoring video. Subsequently, these two types of monitoring video data are transmitted back to the designated receiving end. This can enable the operator to intuitively understand the situation of different levels of the water area and assist in water quality monitoring analysis and navigation decision-making. For example, when monitoring a water area near an industrial wastewater discharge outlet, the bottom camera captures the damage of the underwater pipeline and the abnormal reactions of organisms in the polluted water body around, and the network camera captures the pollutants floating on the water surface and the surrounding environmental conditions. After transmitting back these videos, the staff can comprehensively evaluate the pollution degree and possible sources of this area based on them.

[0034] As Figure 3 shown, further in order to accurately obtain the current water area boundary and the dynamic trajectories of dynamic obstacles by processing the obstacle azimuth information, a map outlining module is proposed, including: An obstacle boundary point marking sub-module, which is used to mark on the electronic map based on the direction and distance in the azimuth information of all environmental obstacles obtained at each moment within a preset period and the self-positioning information of the unmanned ship, and obtain the boundary marking points at each moment within the preset period; The fixed boundary marker sub-module is used to count the marking frequencies of each boundary marker point at all times within a preset period, screen out all boundary marker points with marking frequencies greater than the preset marking frequency threshold among all boundary marker points within the preset period as the fixed boundary points, determine the fixed boundary based on all the fixed boundary points, and regard all the remaining boundary marker points except the fixed boundary points among all the boundary marker points at all times as the first undetermined points; The dynamic trajectory calculation sub-module is used to determine the dynamic trajectories of all dynamic obstacles based on all the first undetermined points and the corresponding calibration times; The water area boundary tracing sub-module is used to determine the current water area boundary based on the dynamic trajectories of all dynamic obstacles and the fixed boundary.

[0035] In this embodiment, marking on the electronic map based on the direction and distance in the azimuth information of all environmental obstacles and the self-positioning information of the unmanned ship obtained at each moment within the preset period to obtain the boundary marker points at each moment within the preset period means that within the set preset period, the sensors of the unmanned ship collect the direction and distance data of all environmental obstacles relative to itself at each moment, and at the same time combine the self-positioning information of the unmanned ship, and then make marks at the corresponding positions on the electronic map. These marked points represent the positions of environmental obstacles relative to the unmanned ship at each moment within the preset period, that is, the boundary marker points. For example, the preset period is 30 minutes, and each 1 minute is taken as a moment. At the 15th minute, the sensor detects an obstacle at a position 40 meters ahead and in the northeast direction. According to the self-positioning information of the unmanned ship, a mark is made at the corresponding position on the electronic map. Continuing this operation, the boundary marker points at each moment within these 30 minutes can be obtained.

[0036] In this embodiment, counting the marking frequencies of each boundary marker point at all times within the preset period means that after the preset period ends, count the number of times each boundary marker point on the electronic map appears in all moments of the entire preset period. This counted number is the marking frequency of the boundary marker point. For example, there are 60 moments within the preset period, and a certain boundary marker point appears in 20 of these moments, then the marking frequency of this boundary marker point is 20 times.

[0037] In this embodiment, the preset marking frequency threshold is a standard value set artificially to judge whether a boundary marking point belongs to a relatively stable boundary. After the marking frequency of each boundary marking point is counted, it will be compared with the preset marking frequency threshold. If the marking frequency of a certain boundary marking point is greater than the threshold, it is inclined to think that the position represented by this point is relatively fixed in the water area and is more likely to be part of the water area boundary. For example, for the monitoring of a specific water area, the preset marking frequency threshold is set to 15 times. If the marking frequency of a boundary marking point reaches 18 times, then it meets this screening condition.

[0038] In this embodiment, determining the determined boundary based on all determined boundary points means comprehensively processing all determined boundary points screened by comparing with the preset marking frequency threshold within a preset period. These determined boundary points appear frequently and relatively stably within the preset period. By connecting and fitting these points, a relatively stable and reliable part of the water area boundary is determined. For example, on an electronic map, there are multiple determined boundary points that meet the condition that the marking frequency is greater than the threshold. They are relatively concentrated and regular in distribution. Connecting these points in sequence, the formed continuous line is determined as the determined boundary, which may be a fixed shoreline in the water area or the boundary around a large fixed obstacle, etc.

[0039] As Figure 3 shown, further, in order to accurately obtain the current water area boundary and the dynamic trajectories of dynamic obstacles by processing the obstacle orientation information, a map outlining module is proposed, including: An obstacle boundary point marking sub-module, which is used to mark on the electronic map based on the direction and distance in the orientation information of all environmental obstacles obtained at each moment within a preset period and the self-positioning information of the unmanned ship, and obtain the boundary marking points at each moment within the preset period; A determined boundary calibration sub-module, which is used to count the marking frequency of each boundary marking point at all moments within a preset period, screen out all boundary marking points with a marking frequency greater than the preset marking frequency threshold among all boundary marking points within the preset period as determined boundary points, determine the determined boundary based on all determined boundary points, and regard all the remaining boundary marking points except the determined boundary points among all boundary marking points at all moments as the first undetermined points; A dynamic trajectory calculation sub-module, which is used to determine the dynamic trajectories of all dynamic obstacles based on all the first undetermined points and the corresponding calibration times; A water area boundary outlining sub-module, which is used to determine the current water area boundary based on the dynamic trajectories of all dynamic obstacles and the determined boundary.

[0040] In this embodiment, the orientation information (including direction and distance) of all environmental obstacles obtained at each moment within a preset period and the self-positioning information of the unmanned ship are marked on an electronic map to obtain boundary marking points at each moment within the preset period. This means that within the set preset period, the sensors of the unmanned ship collect the direction and distance data of all environmental obstacles relative to itself at each moment, and at the same time, combine the self-positioning information of the unmanned ship, and then mark at the corresponding positions on the electronic map. These marked points represent the positions of environmental obstacles relative to the unmanned ship at each moment within the preset period, that is, the boundary marking points. For example, if the preset period is 30 minutes and each 1 minute is regarded as a moment, at the 15th minute, the sensor detects an obstacle at a position 40 meters ahead and in the northeast direction. According to the self-positioning information of the unmanned ship, a mark is made at the corresponding position on the electronic map. By continuously operating in this way, the boundary marking points at each moment within these 30 minutes can be obtained.

[0041] In this embodiment, the marking frequency of each boundary marking point at all moments within the preset period is counted. This means that after the preset period ends, the number of times each boundary marking point appears in all moments of the entire preset period on the electronic map is counted, and this counted number is the marking frequency of the boundary marking point. For example, if there are 60 moments in the preset period and a certain boundary marking point appears in 20 of these moments, then the marking frequency of this boundary marking point is 20 times.

[0042] In this embodiment, a preset marking frequency threshold is a standard value set artificially to judge whether a boundary marking point belongs to a relatively stable boundary. After counting the marking frequency of each boundary marking point, it will be compared with the preset marking frequency threshold. If the marking frequency of a certain boundary marking point is greater than this threshold, it is inclined to be considered that the position represented by this point is relatively fixed in the water area and is more likely to be a part of the water area boundary. For example, for the monitoring of a specific water area, the preset marking frequency threshold is set to 15 times. If the marking frequency of a boundary marking point reaches 18 times, then it meets this screening condition.

[0043] In this embodiment, the determined boundary is determined based on all the determined boundary points. This means that within the preset period, all the determined boundary points screened by comparing with the preset marking frequency threshold are comprehensively processed. These determined boundary points appear frequently and relatively stably within the preset period. By connecting and fitting these points, etc., a relatively stable and reliable part of the water area boundary is determined. For example, on the electronic map, there are multiple determined boundary points that meet the condition that the marking frequency is greater than the threshold. They are relatively concentrated and regular in distribution. Connecting these points in sequence, the formed continuous line is determined as the determined boundary, which may be a fixed shoreline in the water area or the boundary around a large fixed obstacle, etc.

[0044] Further, in order to quantitatively analyze the relative position relationship between the second assumed target drift path and the determined boundary from different perspectives and provide a basis for trajectory rationality evaluation, a relative position analysis unit is proposed, including: A boundary distance factor determination subunit, configured to fit the path line of each second assumed target drift path, determine the tangent line at each position point on each path line in the corresponding path route, and use the intersection point of the line segment passing through each position point on each path line and perpendicular to the tangent line at the corresponding position point with the determined boundary as the distance statistical point of the corresponding position point. Based on the spacing between all position points on each path line and the corresponding distance statistical points, the boundary distance factor of the corresponding second assumed target drift path is determined; A crossing number penalty factor determination subunit, configured to define the calculation formula of the crossing number penalty factor based on an exponential function, and determine the crossing number penalty factor of each second assumed target drift path based on the intersection number of each second assumed target drift path and the determined boundary and the calculation formula of the crossing number penalty factor; A direction consistency factor determination subunit, configured to generate a path tangent angle vector for each second assumed target drift path based on the tangent angles of all position points on the path line of each second assumed target drift path in the corresponding path line. At the same time, a boundary tangent vector for each second assumed target drift path is generated based on the tangent angles of the determined boundary at the distance statistical points of all position points on the path line of each second assumed target drift path. The (cosine) similarity between the path tangent angle vector and the boundary tangent vector of each second assumed target drift path is used as the direction consistency factor.

[0045] In this embodiment, fitting the path line of each second assumed target drift path means using a mathematical fitting method for each discrete point (such as the first undetermined point) that constitutes the second assumed target drift path to construct a continuous curve or broken line to approximately represent the drift path. In this way, a line that can reflect the approximate trajectory of the target drift can be obtained for subsequent in-depth analysis. For example, for multiple discrete points distributed on a certain second assumed target drift path, the polynomial fitting method is used to find a suitable polynomial curve that is as close as possible to these discrete points, and this curve is the fitted path line.

[0046] In this embodiment, based on the distances between all position points and the corresponding distance statistical points on each path line, the boundary distance factor corresponding to the second assumed target drift path is determined. It means that on the path line of the second assumed target drift path obtained by fitting, for each position point, the corresponding distance statistical point is found (for example, the intersection point of the line segment perpendicular to the tangent of the path line passing through each position point and the established boundary), and then the distance between this position point and the distance statistical point is calculated. By comprehensively considering the distances between all these position points on the path line and the corresponding distance statistical points (such as calculating the average value, weighted average value, etc.), a value that can reflect the distance between this path line and the established boundary is obtained, and this value is the boundary distance factor. For example, if there are 10 position points on the path line, and the distances between them and the corresponding distance statistical points are calculated as 2 meters, 3 meters, 2.5 meters... respectively, the average value obtained by averaging these distance values is used as the boundary distance factor to measure the distance relationship between this second assumed target drift path and the established boundary.

[0047] In this embodiment, based on the exponential function, the calculation formula for the crossing - times penalty factor is defined, and based on the number of intersections of each second assumed target drift path with the established boundary and the calculation formula for the crossing - times penalty factor, the crossing - times penalty factor of each second assumed target drift path is determined. It means that first, a formula is set in the form of an exponential function to quantify the influence degree of the number of intersections of the second assumed target drift path with the established boundary on its rationality. Generally speaking, the more crossing times there are, the greater the penalty. For example, the calculation formula for the crossing - times penalty factor is defined as f(n)=a n (where n is the number of crossings, a > 1 is a constant, such as a = 1.5). Then, the actual number of intersections of each second assumed target drift path with the established boundary is counted, and this number of intersections is substituted into this formula for calculation. The result obtained is the crossing - times penalty factor of this second assumed target drift path. For example, if a certain second assumed target drift path intersects the established boundary 3 times, substituting it into the above formula f(3)=1.5^3 = 3.375, then 3.375 is the crossing - times penalty factor of this path. The larger this factor is, the greater the penalty this path receives in the evaluation due to crossing the established boundary.

[0048] Furthermore, in order to accurately determine the current water area boundary, making full use of the boundary marker point information and avoiding the dynamic obstacle trajectories, a water area boundary stroking sub - module is proposed, including: An undefined - point definition unit, which is used to regard all the remaining boundary marker points among all the boundary marker points at all times except those covered by the established boundary and the dynamic trajectories of all dynamic obstacles as the second undefined points; A water area boundary stroking unit is used to perform clustering analysis on second undetermined points with the distance between all second undetermined points as the clustering condition, determine multiple clustering position points, and connect the determined boundary with all clustering position points on the premise of avoiding intersection with the dynamic trajectories of all dynamic obstacles to obtain the current water area boundary.

[0049] In this embodiment, performing clustering analysis on second undetermined points with the distance between all second undetermined points as the clustering condition, determining multiple clustering position points, and connecting the determined boundary with all clustering position points on the premise of avoiding intersection with the dynamic trajectories of all dynamic obstacles to obtain the current water area boundary means that when dealing with the problem of determining the water area boundary, the remaining boundary marker points (i.e., second undetermined points) that neither belong to the determined boundary nor are within the coverage range of the dynamic trajectories of dynamic obstacles are grouped and classified according to their distance relationships. Specifically, calculate the average distance between all pairs of second undetermined points in each clustering combination in the clustering result, and take the average of the average distances of all clustering combinations in a single clustering result as the evaluation value of the clustering result. When the evaluation value of the clustering result converges during the clustering process, take the position point corresponding to the average coordinates of all second undetermined points in each clustering combination in the clustering result at this time as a single clustering position point, so as to obtain multiple clustering position points.

[0050] Furthermore, in order to obtain an accurate current water area boundary, when connecting the determined boundary with these clustering position points, it is necessary to ensure that the connected line does not intersect with the dynamic trajectories of dynamic obstacles, because the movement range of dynamic obstacles may affect the definition of the water area boundary, and the unmanned ship needs to avoid these dynamic obstacles. The line formed in this way is determined as the current water area boundary. For example, the determined boundary is in an L shape, there are multiple clustering position points near the determined boundary, and there is a dynamic trajectory of a dynamic obstacle that obliquely passes through part of the clustering position point area. When connecting the determined boundary with the clustering position points, select a path that does not intersect with this dynamic trajectory for connection, and finally form the current water area boundary.

[0051] Furthermore, in order to comprehensively consider various factors to analyze the mutation reasons of the local mutation boundaries of each type of water quality monitoring data and deduce the local mutation range, the process of the path optimization module determining the local mutation range of all types of water quality monitoring data obtained within a preset period includes: Based on the spatial correlation constraint effect between the current local mutation boundary and the current water area boundary of all types of water quality monitoring data obtained within a preset period, the synergy effect between all types of water quality monitoring data, the trajectory correlation effect of dynamic obstacles, and the preset pollution source mutation mode library, analyze the mutation cause of the current local mutation boundary of each type of water quality monitoring data; Based on the mutation cause tracing of the current local mutation boundary of each type of water quality monitoring data and the current water area boundary, the current local mutation boundary of each type of water quality monitoring data is continuously calculated to obtain the local mutation range of each type of water quality monitoring data.

[0052] In this embodiment, based on the spatial association constraint effect between the current local mutation boundary and the current water area boundary of all types of water quality monitoring data obtained within a preset period, the synergy effect between all types of water quality monitoring data, the trajectory association effect of dynamic obstacles, and the preset pollution source mutation mode library, analyzing the mutation cause tracing of the current local mutation boundary of each type of water quality monitoring data means exploring the reasons for the generation of the local mutation boundary in water quality monitoring data by integrating multiple factors. The spatial association constraint effect considers the connection between the current local mutation boundary and the current water area boundary in terms of spatial position, such as whether the local mutation boundary is close to specific areas of the water area (such as the water inlet, sewage outlet, etc.); the synergy effect focuses on the mutual influence between all types of water quality monitoring data, and the mutation of one water quality index may be associated with the changes of other indicators; the trajectory association effect of dynamic obstacles analyzes whether there is an association between the movement trajectory of dynamic obstacles and the water quality mutation area, for example, whether the dynamic obstacles carry pollutants to affect the water quality; the preset pollution source mutation mode library is the information of various pollution sources and mutation modes that may cause water quality mutation collected and sorted in advance. By comprehensively analyzing these factors, the reasons for the mutation of the current local mutation boundary of each type of water quality monitoring data are found. For example, within a certain preset period, it is found that a local mutation boundary appears in a corner of a certain type of water quality monitoring data near the water area, and there are often ships (dynamic obstacles) transporting chemicals passing by in this area, and there are also certain changes in other related water quality indicators. Combining with the mutation mode in the preset pollution source mutation mode library about chemical leakage causing water quality mutation, it is analyzed that the mutation cause tracing of this local mutation boundary may be chemical leakage during the ship transportation process.

[0053] In this embodiment, based on the mutation cause tracing of the current local mutation boundary of each type of water quality monitoring data and the current water area boundary, the current local mutation boundary of each type of water quality monitoring data is continuously calculated to obtain the local mutation range of each type of water quality monitoring data means that after clarifying the mutation reason, based on the current water area boundary information, the possible affected range of water quality mutation is speculated. Because the mutation cause tracing can prompt potential information such as the propagation direction and speed of water quality mutation, and the current water area boundary limits the spatial range where the mutation may spread. For example, if it is determined that the mutation cause tracing is the sewage discharge from a factory at a fixed location and the water flow direction is known, combined with the current water area boundary (such as the boundary range on both sides of the river), it is possible to calculate the area that may be affected by the local mutation of this type of water quality monitoring data along the water flow direction and considering the limitation of the water area boundary, so as to determine the local mutation range, which may be an area within the range limited by the water area boundary along the water flow direction starting from the factory sewage outlet.

[0054] Furthermore, in order to generate an objective function considering pollution coverage, voyage cost, and obstacle avoidance safety distance and thereby generate an optimized sampling path in real time, a path optimization module is proposed, including: An objective function generation sub-module for generating an objective function based on pollution coverage, voyage cost, and obstacle avoidance safety distance; A sampling path optimization sub-module for generating an optimized sampling path in real time with the goal of minimizing the output value of the objective function.

[0055] In this embodiment, generating an objective function based on pollution coverage, voyage cost, and obstacle avoidance safety distance means constructing a mathematical function by comprehensively considering these three key factors. Pollution coverage represents the coverage degree of the sampling path of the unmanned ship for the possible polluted areas. The higher the coverage degree, the more conducive it is to comprehensively monitor the water quality pollution situation; the voyage cost involves the costs such as energy and time consumed by the unmanned ship when traveling along the sampling path. The shorter the voyage, the lower the cost; the obstacle avoidance safety distance is to ensure that the unmanned ship maintains a safe interval from environmental obstacles during navigation to avoid collisions. By assigning corresponding weight coefficients to these three factors and combining them in a specific mathematical operation method, an objective function is formed. For example, let the pollution coverage be C, the voyage cost be D, and the obstacle avoidance safety distance be S. The objective function may be expressed as F = w1×(1 / C)+w2×D+w3×(1 / S), where w1, w2, and w3 are weight coefficients, and the importance of each factor in the objective function is balanced by adjusting the weight coefficients.

[0056] In this embodiment, generating an optimized sampling path in real time with the goal of minimizing the output value of the objective function means that when generating the sampling path of the unmanned ship, calculations and planning are carried out with the goal of making the value output by the above objective function reach the minimum. Because the objective function comprehensively considers pollution coverage, voyage cost, and obstacle avoidance safety distance, when the value of the objective function is the smallest, it means that on the premise of meeting the obstacle avoidance safety distance, the pollution coverage is improved as much as possible while the voyage cost is reduced, thereby realizing the optimization of the sampling path. For example, using relevant path planning algorithms (such as the A* algorithm, etc.), combined with the real-time obtained environmental information (such as the position of obstacles, the range of local mutations of water quality data, etc.), continuously adjust the planning of the sampling path to make the value of the objective function F gradually decrease. The finally determined path is the optimized sampling path that meets the conditions at the current moment to ensure that the unmanned ship can complete the water quality monitoring task efficiently and safely.

[0057] The present invention provides an implementation manner of a water quality monitoring method based on multi-sensor fusion, including: Based on multiple sensors, the azimuth information of all environmental obstacles and various types of water quality monitoring data when the unmanned ship sails on the water area according to the initial sampling path are obtained in real time; Locate the orientation information of all environmental obstacles obtained within a preset period in chronological order and outline them on an electronic map to obtain the current water area boundary and the dynamic trajectories of dynamic obstacles; Based on the local mutation range of all types of water quality monitoring data obtained within a preset period, as well as the current water area boundary and the dynamic trajectories of dynamic obstacles, generate an optimized sampling path in real time; Control the unmanned ship to navigate on the water area according to the optimized sampling path and transmit back the multi-type water quality monitoring data collected in real time using a variety of sensors.

[0058] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the present invention and its equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. An unmanned ship system for water quality monitoring based on multi-sensor fusion, characterized in that Including: An initial monitoring module, configured to obtain in real time the azimuth information of all environmental obstacles and various types of water quality monitoring data when the unmanned ship sails on the water surface along the initial sampling path based on multiple sensors; A map outlining module, configured to perform chronological positioning on the azimuth information of all environmental obstacles obtained within a preset period and outline them in an electronic map to obtain the current water area boundary and the dynamic trajectories of dynamic obstacles; A path optimization module, configured to generate an optimized sampling path in real time based on the local mutation range of all types of water quality monitoring data obtained within a preset period and the current water area boundary and the dynamic trajectories of dynamic obstacles; A monitoring and transmission module, configured to control the unmanned ship to sail on the water surface along the optimized sampling path and transmit back various types of water quality monitoring data collected in real time by using multiple sensors.

2. The water quality monitoring unmanned ship system based on multi-sensor fusion according to claim 1, characterized in that, The initial monitoring module includes: An obstacle detection sub-module, configured to obtain in real time the distances of all environmental obstacles when the unmanned ship sails on the water surface along the initial sampling path based on a lidar, and obtain the self-positioning information of the unmanned ship when it sails on the water surface along the initial sampling path based on a GPS module; A water quality monitoring sub-module, configured to obtain various types of water quality monitoring data when the unmanned ship sails on the water surface along the initial sampling path based on various types of water quality sensors.

3. The water quality monitoring unmanned ship system based on multi-sensor fusion according to claim 1, characterized in that, It further includes: An underwater camera and a network camera are respectively configured to obtain an underwater monitoring video and an above-water monitoring video and transmit them back.

4. The water quality monitoring unmanned ship system based on multi-sensor fusion according to claim 1, characterized in that, The map outlining module includes: An obstacle boundary point marking sub-module, configured to mark in the electronic map based on the direction and distance in the azimuth information of all environmental obstacles obtained at each moment within a preset period and the self-positioning information of the unmanned ship to obtain boundary marking points at each moment within the preset period; A determined boundary calibration sub-module, configured to count the marking frequencies of each boundary marking point at all moments within a preset period, screen out all boundary marking points with marking frequencies greater than a preset marking frequency threshold among all boundary marking points within the preset period as determined boundary points, determine the determined boundary based on all determined boundary points, and regard all remaining boundary marking points except the determined boundary points among all boundary marking points at all moments as first undetermined points; A dynamic trajectory calculation sub-module, configured to determine the dynamic trajectories of all dynamic obstacles based on all first undetermined points and the corresponding calibration times; A water area boundary outlining sub-module, configured to determine the current water area boundary based on the dynamic trajectories of all dynamic obstacles and the determined boundary.

5. The water quality monitoring unmanned ship system based on multi-sensor fusion according to claim 4, characterized in that, The dynamic trajectory calculation sub-module includes: A hypothesis path calibration unit, configured to summarize all first undetermined points at each moment within a preset period to obtain a set of first undetermined points at each moment, select an arbitrary first undetermined point from the set of first undetermined points at each moment within the preset period and summarize them to obtain all first hypothesis target drift paths, and regard the hypothesis target drift paths that intersect with the determined boundary at most once among all first hypothesis target drift paths as second hypothesis target drift paths; A speed stability evaluation unit, configured to calculate the drift speed stability of each second hypothesis target drift path; A relative position analysis unit, configured to determine a boundary distance factor, a crossing number penalty factor, and a direction consistency factor of each second hypothetical target drift path; an occurrence probability evaluation unit, configured to evaluate the occurrence probability of each hypothetical target drift path based on a boundary distance factor, a crossing number penalty factor, and a direction consistency factor of each second hypothetical target drift path; The rationality determination unit is used to calculate the rationality of each assumed target drift probability based on the drift speed stability and occurrence probability of each second assumed target drift path, and regard all second assumed target drift paths with a rationality greater than a preset rationality threshold as the dynamic trajectories of all dynamic obstacles.

6. The water quality monitoring unmanned ship system based on multi-sensor fusion according to claim 5, characterized in that Relative position analysis unit, including: a boundary distance factor determination subunit, configured to fit a path line of each second hypothetical target drift path, and determine a boundary distance factor corresponding to the second hypothetical target drift path based on the distances between all position points and corresponding distance statistical points in each path line; a crossing number penalty factor determination subunit, configured to define a calculation formula for the crossing number penalty factor based on an exponential function, and determine the crossing number penalty factor for each second hypothetical target drift path based on the number of intersections between each second hypothetical target drift path and the predetermined boundary and the calculation formula for the crossing number penalty factor; The direction consistency factor determination subunit is used to take the similarity between the path tangent angle vector and the boundary tangent vector of each second hypothetical target drift path as the direction consistency factor.

7. The water quality monitoring unmanned ship system based on multi-sensor fusion according to claim 4, characterized in that, Water boundary stroke submodule, including: An undetermined point definition unit is used to treat all boundary marking points at all times, except for all boundary marking points covered by the dynamic trajectories of the determined boundary and all dynamic obstacles, as second undetermined points; The water area boundary stroke unit is used to perform cluster analysis on the second undetermined points based on the distance between all the second undetermined points as the clustering condition, determine multiple cluster location points, and connect the determined boundary with all cluster location points to obtain the current water area boundary while avoiding intersection with the dynamic trajectory of all dynamic obstacles.

8. The water quality monitoring unmanned ship system based on multi-sensor fusion according to claim 1, characterized in that, The path optimization module determines the local mutation range of all types of water quality monitoring data obtained within a preset period, including: Based on the spatial correlation constraint effect between the current local mutation boundary and the current water area boundary of all types of water quality monitoring data obtained within the preset period, the synergistic effect between all types of water quality monitoring data, the trajectory correlation effect of dynamic obstacles, and the preset pollution source mutation pattern library, the mutation traceability of the current local mutation boundary of each type of water quality monitoring data is analyzed; Based on the mutation tracing of the current local mutation boundary of each type of water quality monitoring data and the current water area boundary, the current local mutation boundary of each type of water quality monitoring data is continuously extrapolated to obtain the local mutation range of each type of water quality monitoring data.

9. The water quality monitoring unmanned ship system based on multi-sensor fusion according to claim 1, characterized in that, Path optimization module, including: The objective function generation submodule is used to generate the objective function based on pollution coverage, range cost, and obstacle avoidance safety distance; The sampling path optimization submodule is used to generate an optimized sampling path in real time with the goal of minimizing the output value of the objective function.

10. A water quality monitoring method based on multi-sensor fusion, characterized in that, include: Based on multiple sensors, all the azimuth information of environmental obstacles and various types of water quality monitoring data are obtained in real time when the unmanned ship sails on the water area according to the initial sampling path; The azimuth information of all environmental obstacles obtained within a preset period is located in chronological order and outlined in the electronic map to obtain the current water area boundary and the dynamic trajectory of dynamic obstacles; Based on the local mutation range of all types of water quality monitoring data obtained within a preset period, and the current water area boundary and the dynamic trajectory of dynamic obstacles, an optimized sampling path is generated in real time; Control the unmanned ship to sail on the water area according to the optimized sampling path, and transmit back various types of water quality monitoring data collected in real time by using multiple sensors.

Images